cmp.pkg

d 'CMP': module to implement conditional (recursive) mixed process estimator
d
d  cmp estimates multi-equation, mixed process models, potentially
d with hierarchical random effects. "Mixed process" means that
d different equations can have different kinds of dependent
d variables. The choices are: continuous (like OLS), tobit (left-,
d right-, or bi-censored), probit, ordered probit or fractional
d probit. "Conditional" means that the model can vary by
d observation. An equation can be dropped for observations for
d which it is not relevant--if, say, a worker retraining program is
d not offered in a city then the determinants of uptake cannot be
d modeled there. Or the type of dependent variable can vary by
d observation. A dependent variable in one equation can appear on
d the right side of another equation. Such dependencies must have a
d recursive structure if the dependencies are on censored variables
d as observed, meaning that they split the equations into stages.
d If the dependencies are on (latent) linear dependent variables,
d they can be recursive or simultaneous in structure. So cmp can
d fit many SUR, simultaneous equation, and IV models. cmp's
d modeling framework therefore embraces those of the official Stata
d commands  probit, ivprobit, treatreg, biprobit, tetrachoric,
d oprobit, mprobit, asmprobit, asroprobit, tobit, ivtobit, cnreg,
d intreg, truncreg, heckman, heckprob, xtreg, xtprobit, xttobit,
d xtintreg, in principle even regress, sureg, and reg3. It goes
d beyond them in offering far more flexibility in model
d construction. The routine runs under Stata 11 or later. 
d
d KW: conditional models
d KW: mixed process modules
d KW: probit
d KW: tobit
d KW: ivtobit
d KW: biprobit
d KW: multinomial probit
d KW: ordered probit
d KW: truncated regression
d KW: fractional probit
d
d Requires: Stata version 11 and package ghk2 (q.v.)
d
d Author: David Roodman
d Support: email david@@davidroodman.com
d
f cmp.ado
f cmp_p.ado
f cmp.sthlp
f cmp.mata
f lcmp.mlib
f cmp_clear.ado